November 17, 2022 How to Fix Broken Process Automation

There’s a way to fix broken automated processes: orchestrating the various process endpoints or tying multiple processes together. 

Most of us have experienced some form of broken process automation in our daily lives. At best, these experiences are minor annoyances. But, if automation isn’t done right, it can be detrimental to both businesses and customers alike. According to McKinsey, 65% of smaller companies report success with automation, compared with just 55% at large organizations. What’s preventing that number from being higher?

Perhaps a situation like this has happened to you. Sarah orders a couch from a retailer for her home renovation. She confirms that she can pick up the couch at a showroom locally the same day and selects, “Pick up in store” at the end of her transaction. She gets an email confirmation saying her order is being prepped, and will be ready in just a few hours.

Prior to going to the showroom, she calls to check that her couch is ready for pickup, as she never receives a confirmation notification. The contact center employee confirms via the order fulfillment system that she’s sixth in line for pickup, and that she won’t be able to pick up her order until tomorrow. Now, Sarah is not only annoyed, but she may also never shop with this retailer again.

Fortunately, there’s a way to fix what’s broken with automated processes. The key message is about orchestrating the various process endpoints (think: people, systems and devices), or even tying multiple processes together.

Process automation is greater than the sum of its endpoints

The age-old expression, “The whole is greater than the sum of its parts,” applies to process automation, too. Each “part” of an automated process can be thought of as a “process endpoint.” True end-to-end processes span many process endpoints, which may include people (often cross-functional teams), systems (SaaS applications, homegrown software, legacy systems), and physical devices. Each of these endpoints contributes work to complete the process.

The main cause of process automation failure happens when tasks and automations are disconnected. That can lead to broken end-to-end automations that aren’t integrated with each other. In addition, these disconnects can make it hard to gain visibility into processes and KPIs to measure success. Finally, it becomes difficult to change end-to-end processes, since they involve changes across many different endpoints. That makes processes inflexible. As a result, increasing the degree of automation in your organization involves tying together, or orchestrating, these process endpoints.

Tackling complex processes

Another important point is that most processes out there are pretty complicated, even if they seem like they should be simple. In reality, complex processes need to be coordinated based on a certain logic, which rarely involves a simple sequence of steps. To contrast, most complex process logic must be described by advanced workflow patterns. Many advanced workflow patterns involve handling complex business process logic across multiple endpoints or reacting to events.

For example, the furniture store above might have 15 customers place “pick up in store” orders online. That means the store must process 15 order positions to be picked up in stock and remove them from inventory online. What’s even more difficult is that these processes must happen in parallel. That means dynamically coordinating potentially thousands of tasks when customer orders come in.

Complex processes can be challenging in a few different ways. First, aligning technical and non-technical stakeholders can be difficult to achieve. These two groups must come together to design, operate and improve complex end-to-end processes. Second, executing processes across endpoints must happen in a scalable, repeatable way. You never know when transactions may spike or other anomalies may happen. Systems that offer high availability and redundancy across multiple data centers can alleviate these technical challenges.

Getting started with orchestrating your processes

Process automation, for most companies, shouldn’t be a one-and-done effort. It should be a lifecycle. First, you might take stock of your current processes. What people, devices, and systems do you have in place? One option is to take an inventory of your current processes. Another is to start with a small pilot project as a Proof of Concept (PoC) to test feasibility.

As described above, you may need to orchestrate many different process endpoints. These could include legacy systems, RPA bots, human tasks, microservices, APIs and more. It’s important to view these endpoints holistically, instead of in a silo. Gradually, you can integrate these process endpoints together bit by bit. That is, until you can dedicate the resources to rewriting legacy applications. This type of gradual digital transformation doesn’t have to disrupt your day-to-day business operations.

To summarize, orchestrating your processes makes the most sense when your business processes span a diverse set of process endpoints, and are described by a more complex logic than just a simple sequence of steps. The ideal outcome is fixing what’s broken about automation, and even improving customer loyalty and trust.

Jakob Freund is co-founder and CEO of Camunda – responsible for the company’s vision and strategy. He’s also the driving force behind Camunda’s global growth and takes responsibility for the company culture. As well as holding an MSc in Computer Science, he co-authored the book “Real-Life BPMN” and is a sought-after speaker at technology and industry events.

Subscribe to Industry Today